1. Patrick L Kinney Professor of Environmental Health Sciences Director, Columbia Climate and Health Program plk3@columbia.edu Climate Change, Air Quality and Human Health

2. Overview Health impacts of key air pollutants Ways that climate can affect air pollution Case study 1: health impact assessment for climate change and air quality in NYC metro region Case study 2: climate, pollen, and asthma Future research directions

3. Background The mixtures of air pollutants produced by burning of fuels can –Adversely affect human health –Promote climate change In addition –Climate can influence air pollution, resulting in direct health effects –Climate can affect other aspects of air quality, including smoke from agricultural or wild fires, and aero-allergens like pollen and mold spores

4. Criteria Pollutants: Ubiquitous Pollutants of Human Health Concern Carbon monoxide (CO) Nitrogen dioxide (NO 2 ) Lead (Pb) Sulfur dioxide (SO 2 ) Ozone (O 3 ) Particulate matter (PM 2.5, PM 10 ) “National Ambient Air Quality Standards (NAAQS)”

5. Ozone Main pollutant responsible for photochemical smog, formed via reactions in the atmosphere from primary pollutants (NO x and VOCs) in the presence of sunlight Higher temperatures favor ozone formation Strong oxidants that damage cells lining the respiratory system, resulting in a variety of adverse health outcomes, including lung function decrease, asthma attacks, and premature deaths Ozone is also a greenhouse gas that contributes to warming

6. Source: US EPA (1991); in Kleinman and Lipfert, 1996. Note threshold~90°F (32°C)

7. Particulate Matter (PM 2.5, PM 10 ) Can be either primary or secondary; produced by combustion, atmospheric reactions, and mechanical processes Wide range of physical/chemical properties Wide range of human health impacts, including premature deaths Higher temperatures may favor secondary formation Some particle types contribute to climate warming; others to climate cooling

8. Fine Particle Composition Annual average fine particle data for 2001 from the Look Rock station of the Tennessee Valley Authority. Source: http://www.tva.gov/environment/air/ontheair/fine_particles_smokies.htm

9. Health Effects of Air Pollution Historical experience provides strong evidence for causal relationship between air pollution and premature death Modern epidemiology studies have consistently found significant associations Two primary epidemiologic study designs: –Time series studies of acute effects –Cohort or cross-section studies of chronic effects

10. London Killer Fog, December, 1952 Date Source: UK Met Office http://www.metoffice.gov.uk/ education/secondary/students/smog.html

11. Acute Mortality Responses to PM in US, Europe, and Asia Exposure Risks 0.46 0.62 0.5 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 US (90 Cities)*Eur (21 Studies)* Asia (6 Studies) Percent Increase % Change per 10 ug/m3 0.46 0.62 0.5 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 US (90 Cities)*Eur (21 Studies)* Asia (6 Studies) Percent Increase Source: Greenbaum and O’Keefe, BAQ 2003

12. Ozone & Acute Deaths Bell et al., JAMA 2004

13. Long-term average concentrations of fine particle air pollution were associated with mortality rates, controlling for individual-level risk factors across six US cities Results from Harvard Six Cities Study (Dockery et al., NEJM, 1993)

14. Health Effects Summary Outdoor levels of six air pollutants are regulated in the U.S. based on national standards that are assumed to avoid adverse health impacts Ozone and fine particles dominate current health concerns Many health effects have been associated with currently- observed distributions of ozone and fine particles, but mortality is most prominent among them Mortality impacts for fine particles tend to dominate most health impact assessments

15. How might Climate Change Affect Air Pollution? Formation reactions for secondary pollutants generally happen faster at high temp and with greater sunlight Biogenic precursor emissions increase at higher temp Some particle species may volatilize at higher temperatures (e.g. nitrates, organics) Regional air mass patterns over time and space may change, altering stagnation and clearance events The mixing height of the lower atmosphere may change, affecting dilution of pollution emitted at the surface  Use Coupled Climate/Air Quality Models to Investigate

16. Case Study 1: Integrated health impact assessment for climate change and air quality in NYC metro region Linking models for global and regional climate, land use and cover, and air quality, examine the potential public health impacts of heat and air pollution under alternative scenarios of climate change and regional land use in the 2020s, 2050s, and 2080s in the NYC metropolitan region. …grew out of MEC regional assessment (Rosenzweig, Solecki et al)

17. Approach Develop an integrated modeling system that links changes in global climate, regional climate, land use, and emissions, to project future ozone and heat at policy-relevant geographic scales Develop exposure-response functions for temperature and ozone using historical data from the NYC metro area Run the models using alternative greenhouse gas growth scenarios Combine to assess potential mortality risks in the NYC metro area in the 21st century Funded by USEPA STAR grant

18. Integrated Modeling System Global Climate NASA-GISS GCM Regional Climate MM5 Air Quality SMOKE, CMAQ Public Health Risk Assessment Changing Regional Land Use / Land Cover SLEUTH, Remote Sensing, IPCC SRES Scenarios Changing Greenhouse Gas Emissions IPCC SRES Scenarios Changing Ozone Precursor Emissions IPCC SRES Scenarios

19. The Study Team: Health: Patrick Kinney (PI), Joyce Rosenthal, Kim Knowlton, Mailman School of Public Health, Columbia University; NRDC Climate: Cynthia Rosenzweig, Richard Goldberg, Barry Lynn, NASA-Goddard Institute for Space Studies; David Werth & Roni Avissar, Duke University Land use/remote sensing: William Solecki, Jennifer Cox, Hunter College Geography Dept; Christopher Small, Lamont Doherty Earth Observatory; Air quality: Christian Hogrefe, SUNY Albany; Michael Ku, Kevin Civerolo, NY State Dept Environ Conservation; Tracey Holloway, Earth Institute and University of Wisconsin-Madison; Michelle Bell, Yale

20. Knowlton_ICUC5_2003 Downscaling of projections was a key objective  Scale: global (4°x5°) vs. regional (36 km)?  How to project regional & local health impacts? 4 ° x 5 ° lat/long 36 km grid Source: C Rosenzweig & R Goldberg, NASA/GISS

21. Model Setup GISS coupled global ocean/atmosphere model driven by IPCC greenhouse gas scenarios (“A2” and “B2”) MM5 regional climate model took initial and boundary conditions from GISS GCM, and run on 2 nested domains of 108 km and 36 km over the U.S. CMAQ is run at 36 km to simulate ozone 1996 U.S. Emissions processed by SMOKE and – for some simulations - scaled by IPCC scenarios Simulations periods : June – August 1993-1997 June – August 2023-2027 June – August 2053-2057 June – August 2083-2087

23. Hogrefe et al., J. Geophysical Research, 2004 ∆2080s∆2050s 1990s∆2020s

24. Develop exposure-response functions for temperature and ozone using historical data from the NYC metro area Model Inputs Model Outputs OUTCOME: All Internal-Cause Daily Deaths at County Level (JJA: 1990-1999) POISSON Regression PREDICTOR: Daily Ozone from 16 stations PREDICTOR: Daily mean Temp. from 16 stations Day of Week (Indicator Variable); Spline of time β Coefficient Estimates (Standard Errors) Input to Risk Assessment Final Model: log (daily deaths)= DOW + spline(time)+ b1(mean Tlag0)1-3 + b2(max O3 lag0-1)

26. Study Population x Base Rate x Change in Environmental Conditions x Concentration Response Function (CRF) = Number of Additional Deaths Year 2000 county population Baseline county mortality rate Projected from Integrated Model:  temp,  O 3 % increase in mortality per unit  Health Impact Assessment

27. Knowlton et al., Environ Health Perspec, 2004 Modeled changes in: Mean 1-hr max O3 (ppb) O3-related deaths (%)

28. Upper tail of ozone distribution is more sensitive to climate Frequency distribution of the simulated daily ozone maxima averaged over southern Germany during summer (June-August) for the years 1991-2000 and 2031-2039. Right side: zoom of the high-ozone portion of the curve. From Forkel and Knoche 2006.

29. Ongoing work: Examining the distribution of ozone and heat extremes at the county level across the entire 36 km domain.

30. Case Study Summary Health impact assessment was carried out to examine effects of climate change on air quality and resulting human health Holding ozone precursor emissions constant, increased ozone-related acute mortality was modeled under 2020s, 2050s, and 2080s climate Multiple uncertainties should be kept in mind

31. Beyond climate, future air pollution concentrations will depend on: Anthropogenic Emissions Intercontinental Transport Land Use Furthermore, future pollution-related health impacts will depend on future population, age, health status, etc.

32. Case Study 2 Climate, Pollen and Asthma Beggs and Bambrick, EHP 2005

33. Start date of Birch Pollen Season in Brussels 1970-2006: Days after Jan 1st (5-yr running means) Emberlin et al., Int J Biomet, 2002

36. Next steps: weather, pollen, asthma Quantify the exposure-response relationships linking pollen and health outcomes, including asthma emergency department visits, and allergy medication sales Analyze the effects of seasonal and daily meteorology on timing and magnitude of pollen peaks Examine future climate scenarios and their potential impacts on pollen and allergic asthma responses

37. From: Singer et al., Functional Plant Biology 2005, 32, 667-670. Ragweed allergen production increases as a function of CO 2 concentration Further:

38. Atmospheric CO 2 estimated from the 14C content of  celulose extracted from tree cores at each of the sites from the urban to rural gradient. (unpublished data of D. Hsueh & K.L. Griffin).

39. Next steps: tree pollen production and immune response across the CO2 gradient Demonstrate the feasibility of quantitative tree pollen collection from three sampling sites that range from low to high in average CO 2 concentrations and temperatures using Red Oak as a sentinel species. Analyze whether red oak pollen production increases (mass per flower cluster) with increasing CO 2 and temperature. Test whether the intensity of human allergic immune response (as measured by tree pollen extract-induced lymphocyte proliferation, differentiation and cytokine production) increases across the rural to urban gradient.

40. Air Quality and Climate: Future Research Directions There is a need for multi-model, multi-center ensemble studies using global-to-regional climate and air pollution model systems Extend ozone health impact work to Europe Examine future changes in fine particulate matter Analyze the health co-benefits of alternative greenhouse pollutant mitigation strategies  Study further the relationships among weather, CO2, pollen, allergy and asthma

41. Collaborators Cynthia Rosenzweig – NASA GISS Christian Hogrefe – New York Dept of Environmental Conservation Kim Knowlton – Natural Resources Defense Council Perry Sheffield – Mt. Sinai School of Medicine Tom Matte – New York City Dept. of Health Kaz Ito – New York University Kate Weinberger – Columbia University

42. Merci! STRATUS CONSULTING